CN115495954A - Unequal-volume data modeling method and device for combine harvester - Google Patents
Unequal-volume data modeling method and device for combine harvester Download PDFInfo
- Publication number
- CN115495954A CN115495954A CN202211200762.7A CN202211200762A CN115495954A CN 115495954 A CN115495954 A CN 115495954A CN 202211200762 A CN202211200762 A CN 202211200762A CN 115495954 A CN115495954 A CN 115495954A
- Authority
- CN
- China
- Prior art keywords
- data
- parameter
- unequal
- combine harvester
- sample parameter
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
- 238000000034 method Methods 0.000 title claims abstract description 63
- 238000005457 optimization Methods 0.000 claims abstract description 13
- 238000007781 pre-processing Methods 0.000 claims abstract description 10
- 238000012549 training Methods 0.000 claims description 34
- 241000282461 Canis lupus Species 0.000 claims description 23
- 230000006870 function Effects 0.000 claims description 23
- 239000011159 matrix material Substances 0.000 claims description 23
- 238000004088 simulation Methods 0.000 claims description 13
- 238000012545 processing Methods 0.000 claims description 12
- 238000012544 monitoring process Methods 0.000 claims description 11
- 230000008569 process Effects 0.000 claims description 11
- 238000004422 calculation algorithm Methods 0.000 claims description 9
- 230000003247 decreasing effect Effects 0.000 claims description 4
- 238000006243 chemical reaction Methods 0.000 claims description 3
- 238000012163 sequencing technique Methods 0.000 claims description 3
- 241000894007 species Species 0.000 claims description 3
- 238000013461 design Methods 0.000 abstract description 6
- 238000005516 engineering process Methods 0.000 abstract description 6
- 238000004458 analytical method Methods 0.000 abstract description 3
- 238000012423 maintenance Methods 0.000 abstract description 2
- 238000010586 diagram Methods 0.000 description 6
- 150000001875 compounds Chemical class 0.000 description 5
- 239000000463 material Substances 0.000 description 5
- 238000004140 cleaning Methods 0.000 description 3
- 238000012938 design process Methods 0.000 description 3
- 230000000694 effects Effects 0.000 description 3
- 229910000746 Structural steel Inorganic materials 0.000 description 2
- 235000013339 cereals Nutrition 0.000 description 2
- 238000005259 measurement Methods 0.000 description 2
- 239000000203 mixture Substances 0.000 description 2
- 229910000881 Cu alloy Inorganic materials 0.000 description 1
- 206010063385 Intellectualisation Diseases 0.000 description 1
- 240000007594 Oryza sativa Species 0.000 description 1
- 235000007164 Oryza sativa Nutrition 0.000 description 1
- 229910000831 Steel Inorganic materials 0.000 description 1
- 241000209140 Triticum Species 0.000 description 1
- 235000021307 Triticum Nutrition 0.000 description 1
- 230000009471 action Effects 0.000 description 1
- 229910045601 alloy Inorganic materials 0.000 description 1
- 239000000956 alloy Substances 0.000 description 1
- 238000013459 approach Methods 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 238000004364 calculation method Methods 0.000 description 1
- 238000010276 construction Methods 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 238000013401 experimental design Methods 0.000 description 1
- 238000002474 experimental method Methods 0.000 description 1
- 238000003306 harvesting Methods 0.000 description 1
- 230000003993 interaction Effects 0.000 description 1
- 238000011835 investigation Methods 0.000 description 1
- 238000004519 manufacturing process Methods 0.000 description 1
- 230000007246 mechanism Effects 0.000 description 1
- 238000005065 mining Methods 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 230000001151 other effect Effects 0.000 description 1
- 235000009566 rice Nutrition 0.000 description 1
- 230000003068 static effect Effects 0.000 description 1
- 239000010959 steel Substances 0.000 description 1
- 238000012360 testing method Methods 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F30/00—Computer-aided design [CAD]
- G06F30/20—Design optimisation, verification or simulation
- G06F30/23—Design optimisation, verification or simulation using finite element methods [FEM] or finite difference methods [FDM]
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F30/00—Computer-aided design [CAD]
- G06F30/20—Design optimisation, verification or simulation
- G06F30/27—Design optimisation, verification or simulation using machine learning, e.g. artificial intelligence, neural networks, support vector machines [SVM] or training a model
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/004—Artificial life, i.e. computing arrangements simulating life
- G06N3/006—Artificial life, i.e. computing arrangements simulating life based on simulated virtual individual or collective life forms, e.g. social simulations or particle swarm optimisation [PSO]
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F2111/00—Details relating to CAD techniques
- G06F2111/08—Probabilistic or stochastic CAD
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Evolutionary Computation (AREA)
- General Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- Artificial Intelligence (AREA)
- Computer Hardware Design (AREA)
- Software Systems (AREA)
- Geometry (AREA)
- Biomedical Technology (AREA)
- Life Sciences & Earth Sciences (AREA)
- Biophysics (AREA)
- Computational Linguistics (AREA)
- Data Mining & Analysis (AREA)
- General Health & Medical Sciences (AREA)
- Molecular Biology (AREA)
- Computing Systems (AREA)
- Mathematical Physics (AREA)
- Health & Medical Sciences (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Medical Informatics (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The invention provides an unequal capacity data modeling method and device for a combine harvester, and the method comprises the following steps: s1, acquiring operation data of the combine harvester; s2, preprocessing the work data of the combine harvester obtained in the step S1, and calibrating a large sample parameter, a small sample parameter and a prediction target parameter; s3, constructing an unequal capacity data kriging prediction model, and depicting the correlation between the large sample parameter and the small sample parameter; and S4, optimizing the parameters of the unequal capacity data kriging prediction model in the step S3 to obtain the optimized unequal capacity data kriging prediction model. The invention designs a modeling method suitable for the variable-capacity data of the combine harvester, which can better realize the design, analysis, optimization and operation and maintenance of the data-driven combine harvester and provide model support for the combination of the data-driven prediction technology and the depth of the combine harvester.
Description
Technical Field
The invention belongs to the technical field of big data of a combined harvester, and particularly relates to an unequal capacity data modeling method and device for the combined harvester.
Background
The combine harvester is key agricultural machinery equipment for harvesting main grain crops such as rice, wheat, rape and the like, and the operation efficiency and performance of the combine harvester have important significance on grain production and safety in China.
With the rapid development of the intellectualization of agricultural machinery equipment, the operation monitoring of the combine harvester is improved day by day, and a large amount of measured data is generated and accumulated on the experiment and operation site. These data record not only the operation processes inside the combine, but also the interaction mechanism of the combine with the crop. By mining the internal association information of the data, the association relation between the design, control and operation scheme of the combine harvester and the operation performance can be accurately described, the field operation of the combine harvester is guided, and the operation scheme of the combine harvester is improved and optimized, which is also based on accurate data modeling.
Therefore, the simulation data and the actual measurement data of the combine harvester have clear data modeling requirements. The kriging method is one of data modeling methods widely applied in the engineering field. However, the conventional kriging method assumes that the sample volumes among the data parameters are consistent, but the sample volumes among the data parameters of the combine harvester often have a certain difference, that is, the combine harvester data are not tolerant data, so that the kriging method is difficult to apply and exert the efficiency in the application process. For example, the rotation speed of a threshing cylinder of a certain combine harvester can be adjusted in a stepless way, and the threshing clearance can be selected from 4 gears only. The sample capacity of the threshing cylinder rotation speed is equal to the sample capacity of the data set, the sample capacity of the threshing gap is 4, and the sample capacities are different. The structure size parameters of a certain part of the combine harvester are 3, the structure size parameters can be changed randomly in an upper boundary and a lower boundary, the material types are only 2, static simulation is performed for 30 times in the design process, the sample capacity corresponding to the size parameters is 30, the sample capacity corresponding to the material parameters is 2, and the sample capacities are not equal.
Disclosure of Invention
In view of the above technical problems, one of the objectives of one embodiment of the present invention is to provide a method for modeling unequal capacity data of a combine harvester, which includes obtaining operation data of the combine harvester, preprocessing the operation data, and calibrating a large sample parameter x L Small sample parameter x S And predicting a target parameter y, constructing an unequal tolerance data kriging prediction model, and optimizing the kriging model parameters by adopting a wolf optimization algorithmDescribing a large sample parameter x by constructing a specific covariance matrix L With small sample parameter x S And the correlation provides model support for the combination of the data-driven prediction technology and the depth of the combine harvester.
One of the purposes of one mode of the invention is to provide a combine harvester unequal capacity data modeling device, which comprises a data input module, a data processing module, a data modeling module and a data prediction module, wherein the data input module, the data processing module, the data modeling module and the data prediction module are used for modeling and optimizing by adopting a combine harvester unequal capacity data modeling method, and model support is provided for the deep combination of a data-driven prediction technology and a combine harvester.
Note that the description of these objects does not preclude the existence of other objects. It is not necessary for one embodiment of the invention to achieve all of the above objectives. Objects other than the above objects can be extracted from the descriptions of the specification, drawings, and claims.
The invention realizes the technical purpose through the following technical means:
a method for modeling unequal capacity data of a combined harvester comprises the following steps:
s1, acquiring data of a combine harvester;
s2, preprocessing the data of the combine harvester obtained in the step S1, and calibrating a large sample parameter x L Small sample parameter x S And a predicted target parameter y;
s3, constructing an unequal capacity data kriging prediction model and depicting a large sample parameter x L With small sample parameter x S Correlation relations;
and S4, optimizing the parameters of the unequal tolerance data kriging prediction model in the step S3 to obtain the optimized unequal tolerance data kriging prediction model.
Further, the ways for acquiring the data of the combine harvester in the step S1 comprise numerical simulation, vehicle-mounted monitoring, remote monitoring and an intelligent farm platform.
Further, in the step S3, the sample parameter and the prediction target parameter in the unequal tolerance data kriging prediction model satisfy the following relationship:
wherein y is a predicted target parameter;
x is an input variable comprising a large sample parameter x L Small sample parameter x S ，x＝(x L ,x S )；
f j (x) Is the jth basis function;
β j coefficients for the jth basis function;
p is the number of basis functions;
z (x) is a Gaussian process.
Further, the gaussian process Z (x) in step S3 satisfies the following condition:
E(Z(x))＝0；
E(Z(x i )Z(x j ))＝σ 2 R(θ,x i ,x j )；
in the formula, E (. Cndot.) represents the expectation of the variable, the same applies below;
σ 2 is the sample variance;
R(θ,x i ,x j ) Is a correlation matrix;
theta is a parameter vector of the correlation matrix;
x i is an input vector for the ith data;
x j is the input vector for the jth data.
Further, the correlation matrix R (θ, x) i ,x j ) Characterizing a large sample parameter x L And x of small sample parameter S The association relationship between the two is as follows:
in the formula, σ s Is the variance corresponding to the small sample parameter;
σ l is the variance corresponding to the large sample parameter;
ρ is the large sample parameter x L And x of small sample parameter S The regulation relation between the two;
p is the number of small sample parameters;
q is the number of large sample parameters.
Further, in the step S4, a grayish wolf optimization algorithm is adopted to obtain the optimal parameters of the model, and the parameters sigma to be optimized are calculated s 、
σ l 、
Rho is converted into the position of the ith wolf in the species during hunting
Optimizing, wherein the conversion formula is as follows:
in the above scheme, in step S4, a Root Mean Square Error (RMSE) is used to evaluate the training error:
in the formula, n is the number of training samples;
y i is the true output value of the ith training sample;
Further, the optimization of the gray wolf comprises the following specific steps:
step 1) setting population scale, maximum iteration times and searching space; initializing the population and randomly generating a convergence factor
And perturbation vector
Assuming that the number of iterations g =1,
a vector of random numbers between 0 and 1,
is a random number between 0 and 1,
linearly decreasing from 2 with increasing number of iterationsTo 0;
step 2) calculating and sequencing training errors RMSE corresponding to each individual, and recording the individual position with the minimum training error as the position
Step 3) updating the position of each individual;
step 4) updating decreasing numerical value
Convergence factor
And disturbance vector
A parameter;
step 5) judging whether the algorithm meets the convergence condition, if so, ending and giving a final result
Namely the optimal model parameters; otherwise, let g = g +1, return to step 2).
The optimized unequal tolerance data kriging prediction model in the step S4 is as follows:
f(x new ) Denotes x new A vector composed of values obtained by basis functions;
middle and upper corner mark -1 Expressing inversion, the same is applied below;
middle and upper corner mark T Expressing inversion, the same is applied below;
wherein F is a basis function matrix;
F＝(f(x 1 ),f(x 2 ),…,f(x n )) T ；
n is the number of training samples;
F T is the transpose of F;
y is a vector formed by the output of the training data;
The unequal-capacity data modeling device of the combined harvester comprises a data input module, a data processing module, a data modeling module and a data prediction module.
The data input module is used for inputting unequal capacity data generated by each operation part of the combine harvester;
the data processing module is used for preprocessing the data of the combine harvester and calibrating a large sample parameter x L Small sample parameter x S And a predicted target parameter y;
the data modeling module is used for describing a large sample parameter x L With a small sample parameter x S Constructing an unequal capacity data kriging prediction model according to the correlation, and optimizing parameters of the model to obtain an optimized unequal capacity data kriging prediction model;
and the data prediction module is used for predicting a new sample according to the optimized unequal tolerance data Krigin prediction model obtained by the data modeling module.
Compared with the prior art, the invention has the beneficial effects that:
according to one mode of the invention, the invention provides a method for modeling the unequal capacity data of the combine harvester, which comprises the steps of preprocessing after the data of the combine harvester is obtained, and calibrating the parameter x of a large sample L Small sample parameter x S And predicting a target parameter y, constructing an unequal tolerance data kriging prediction model, optimizing kriging model parameters by adopting a wolf optimization algorithm, and depicting a large sample parameter x by constructing a specific covariance matrix L With a small sample parameter x S And the correlation provides model support for the combination of the data-driven prediction technology and the depth of the combine harvester.
According to one mode of the invention, the unequal-capacity data modeling device for the combine harvester comprises a data input module, a data processing module, a data modeling module and a data prediction module, wherein the unequal-capacity data modeling method for the combine harvester is adopted for modeling and optimizing, and a model support is provided for the depth combination of a data-driven prediction technology and the combine harvester
Note that the description of these effects does not hinder the existence of other effects. One embodiment of the present invention does not necessarily have all the effects described above. Effects other than the above can be clearly seen and extracted from the descriptions of the specification, the drawings, the claims, and the like.
Drawings
FIG. 1 is an algorithmic flow diagram of one embodiment of the invention.
Fig. 2 is a schematic structural diagram of an embodiment of the present invention.
Fig. 3 is a component structure diagram of embodiment 1 of the present invention.
FIG. 4 is a finite element simulation diagram of a part according to embodiment 1 of the present invention.
FIG. 5 is a diagram showing the predicted results in example 1 of the present invention.
FIG. 6 is a diagram illustrating the predicted results of embodiment 2 of the present invention.
Detailed Description
Reference will now be made in detail to embodiments of the present invention, examples of which are illustrated in the accompanying drawings, wherein like or similar reference numerals refer to the same or similar elements or elements having the same or similar function throughout. The embodiments described below with reference to the drawings are illustrative and intended to be illustrative of the invention and are not to be construed as limiting the invention.
In the description of the present invention, it is to be understood that the terms "center", "longitudinal", "lateral", "length", "width", "thickness", "front", "rear", "left", "right", "upper", "lower", "axial", "radial", "vertical", "horizontal", "inner", "outer", etc. indicate orientations and positional relationships based on those shown in the drawings, and are used merely for convenience in describing the present invention and for simplicity in description, and do not indicate or imply that the device or element so referred to must have a particular orientation, be constructed and operated in a particular orientation, and thus should not be considered as limiting the present invention. Furthermore, the terms "first", "second" and "first" are used for descriptive purposes only and are not to be construed as indicating or implying relative importance or implicitly indicating the number of technical features indicated. Thus, a feature defined as "first" or "second" may explicitly or implicitly include one or more of that feature. In the description of the present invention, "a plurality" means two or more unless specifically defined otherwise.
In the present invention, unless otherwise expressly specified or limited, the terms "mounted," "connected," "secured," and the like are to be construed broadly and can, for example, be fixedly connected, detachably connected, or integrally connected; can be mechanically or electrically connected; they may be connected directly or indirectly through intervening media, or they may be interconnected between two elements. The specific meanings of the above terms in the present invention can be understood by those skilled in the art according to specific situations.
FIG. 1 shows a preferred embodiment of the method for modeling the unequal capacity data of the combine harvester,
a method for modeling unequal capacity data of a combined harvester comprises the following steps:
s1, acquiring operation data of the combine harvester;
s2, preprocessing the work data of the combine harvester obtained in the step S1, and calibrating a large sample parameter x L Small sample parameter x S And a predicted target parameter y;
s3, constructing an unequal capacity data kriging prediction model and depicting a large sample parameter x L With small sample parameter x S Correlation relations;
and S4, optimizing the parameters of the unequal tolerance data kriging prediction model in the step S3 to obtain the optimized unequal tolerance data kriging prediction model.
Furthermore, the means for acquiring the work data of the combine harvester in the step S1 comprise numerical simulation, vehicle-mounted monitoring, remote monitoring and an intelligent farm platform.
Further, the sample parameter and the prediction target parameter in the unequal tolerance data kriging prediction model in the step S3 satisfy the following relationship:
wherein y is a predicted target parameter;
x is an input variable comprising a large sample parameter x L Small sample parameter x S ，x＝(x L ,x S )；
f j (x) Is the jth basis function;
β j coefficients for the jth basis function;
p is the number of basis functions;
z (x) is a Gaussian process.
Further, the gaussian process Z (x) in step S3 satisfies the following condition:
E(Z(x))＝0；
E(Z(x i )Z(x j ))＝σ 2 R(θ,x i ,x j )；
wherein E (-) represents the expectation of the variable;
σ 2 is the sample variance;
R(θ,x i ,x j ) Is a correlation matrix;
theta is a parameter vector of the correlation matrix;
x i is an input vector for the ith data;
x j is the input vector for the jth data.
Further, the correlation matrix R (θ, x) i ,x j ) Characterizing a large sample parameter x L And x of small sample parameter S The association relationship between the two is as follows:
in the formula, σ s Is the variance corresponding to the small sample parameter;
σ l is the variance corresponding to the large sample parameter;
p is the number of small sample parameters;
q is the number of large sample parameters;
ρ is the large sample parameter x L And x of small sample parameter S Adjustment coefficient between for adjusting large sample parameter x L And x of small sample parameter S The association relationship between them.
The specific construction steps of the unequal capacity data kriging prediction model are as follows:
constructing a basis function matrix F:
F＝(f(x 1 ),f(z 2 ),…,f(x n )) T
n is the number of samples of training data;
f(x i ) Represents the ithThe data are presented as the values obtained by basis functions, as follows.
f(x i )＝(f 1 (x i ),f 2 (x i ),..,f p (x i ))
The input vector is represented as follows:
Y＝Fβ+Z
beta is a vector consisting of basis function coefficients;
z is the gaussian process of the training data.
Z＝(Z(x 1 ),Z(x 2 ),…,Z(x n )) T
Sample x to be predicted new The output predicted value is:
y is a vector formed by the parameters of the training data prediction target; c (x) T Is the transpose of the coefficient vector.
Sample x new The prediction error is as follows.
y(x new ) Is a sample x new The true output value of (d).
From linear unbiased conditions
F T c(x)-f(x)＝0
Calculating regression mean square error
σ 2 ＝D(Z)＝E(Z 2 )-(E(Z)) 2 ＝E(Z 2 )
And establishing the optimal parameters of the optimization problem solving model as follows.
s.t.F T c(x)-f(z)＝0
A langerhans function is constructed.
λ is the lagrange multiplier.
Calculating a deviation to obtain
To obtain
An estimate of z is obtained by the following equation.
By
Get the best estimate of beta
Sample x new Prediction value
As follows:
further, in the step S4, a grey wolf optimization algorithm is adopted to obtain model optimal parameters, and the parameter sigma to be optimized is calculated s 、
σ l 、
Rho is converted into the position of the ith wolf in the species group in hunting
Optimizing, wherein the conversion formula is as follows:
in the above scheme, in the step S4, a Root Mean Square Error (RMSE) is used to evaluate the training error:
in the formula, n is the number of training samples;
y i is the true output value of the ith training sample;
The grey wolf algorithm is a heuristic optimization algorithm, the grey wolf life is mainly based on the population, and a pyramid type hierarchical structure exists in the population. The highest layer is wolf, which is mainly responsible for the decision of hunting action of the population and is called Alpha; the level 2 grey wolf is responsible for assisting the head wolf in making decisions, called Beta; the 3 rd level gray wolf is called Delta and is responsible for investigation and search; the lowest level wolf is called Omega. In the hunting process, the wolf group searches, tracks and approaches to the prey in a team mode, then surrounds the prey from each orientation, when the surrounding circle is small enough, the wolf head Alpha leads Beta and Delta to kill, other wolfs provide support, and the orientation is continuously changed along with the movement of the prey, so that the prey is attacked, and finally the prey is captured.
The embodiment solves the model parameter optimization problem by simulating the hunting behavior of the wolf population, and optimizes the parameter sigma to be optimized s 、
σ l 、
ρ is converted to the following vector:
setting the number of the gray wolfs as N;
(Vector)
indicating the location of the ith wolf in the hunting population.
And recording the individual with the minimum training error as Alpha, recording the suboptimal and third-best individuals as Beta and Delta respectively, and recording other individuals as Omega.
Calculating the relative distance between individuals by the following formula
In the formula (I), the compound is shown in the specification,
representing the positions of individuals Alpha, beta, delta;
(Vector)
for the perturbation vector, as shown below
In the formula (I), the compound is shown in the specification,
a vector of random numbers between 0 and 1.
Each individual location is updated according to the following formula:
wherein g is the current iteration number;
in the formula (I), the compound is shown in the specification,
is a random number between 0 and 1;
Each individual updates the location according to the location of the Alpha, beta, delta individual as follows:
through the continuous iteration, the final position of Alpha is regarded as the optimal solution.
Further, the optimization of the gray wolf comprises the following specific steps:
step 1) setting population scale, maximum iteration times and search space; initializing a population and randomly generating parameters
And
setting the iteration times g =1;
step 2) calculating and sequencing training errors RMSE corresponding to each individual, and recording the individual position with the minimum training error as the position
Step 3) updating the position of each individual;
step 4) updating
Convergence factor
And disturbance vector
A parameter;
step 5) judging whether the algorithm meets the convergence condition, if so, finishing and giving a final result
Namely the optimal model parameters; otherwise, let g = g +1, return to step 2).
The unequal-capacity data modeling device of the combined harvester comprises a data input module, a data processing module, a data modeling module and a data prediction module.
The data input module is used for inputting unequal capacity data generated by each operation part of the combine harvester;
the data processing module is used for preprocessing the operation data of the combine harvester and calibrating a large sample parameter x L Small sample parameter x S And a predicted target parameter y;
the data modeling module is used for describing a large sample parameter x L With small sample parameter x S Constructing an unequal capacity data kriging prediction model according to the correlation relation, and optimizing parameters of the model to obtain an optimized unequal capacity data kriging prediction model;
and the data prediction module is used for predicting a new sample according to the optimized unequal tolerance data Krigin prediction model obtained by the data modeling module.
Example 1
In the embodiment, the modeling of the simulation unequal capacity data generated in the design process of the combine harvester is realized by using the unequal capacity data modeling method of the combine harvester provided by the invention.
1) Design of experimental protocol
The object of this embodiment is a combine chassis attachment, as shown in fig. 3. The structural dimension parameters of the part comprise h, r 1 ，r 2 The material comprises structural steel and copper alloy.
During the design process, for each material, 15 sets of structural dimension parameter samples were generated using the latin hypercube method, and the set of part design variables 30 (15 × 2) was obtained.
For each sample of the structure dimension parameter, a corresponding three-dimensional model is generated.
2) Finite element simulation
The basic idea of the finite element method is to discretize the structure into a finite number of combinations of elements that are connected to each other in a certain way to simulate or approximate the original structure, thereby reducing a continuous infinite degree of freedom problem into a numerical analysis method for solving the discrete finite degree of freedom problem.
The structure was numerically simulated using Ansys software. And dividing the three-dimensional model by using a Mesh function carried in Ansys software, wherein the number of the finally divided grid-divided units is 3744, and the number of the nodes of the units is 17974.
The part A area is fixedly restrained, the part B area is fixedly restrained, 100N load is applied, the part material is structural steel, the elastic modulus is 200GPa, and the Poisson ratio is 0.30, and the structure steel is shown in figure 3.
The maximum stress of the alloy is calculated by using a finite element method, wherein the maximum stress is 22.213MPa as shown in FIG. 4.
A training data set of sample capacity 30 was formed from the experimental design and the simulated maximum stress obtained.
3) Simulation data modeling
Based on the obtained simulation data, a stress prediction model is respectively constructed by adopting a traditional kriging method and the unequal capacity data modeling method of the combined harvester, wherein the optimal parameters of the method provided by the invention are as follows:
[2.98，0.68，1.17，0.95，0.80，1.02]。
according to the method for modeling the unequal capacity data of the combined harvester, provided by the invention, the data x to be predicted are obtained through the following steps new The predicted output value of (a);
step 1) constructing a correlation matrix of large sample parameters and small sample parameters based on the optimal parameters;
in the formula, σ s Is the variance corresponding to the small sample parameter, which is 2.98;
σ l is the variance corresponding to the large sample parameter, which is 1.17;
ρ is the large sample parameter x L And x of small sample parameter S Adjustment coefficient of between, for adjusting a large sample parameter x L And x of small sample parameter S The correlation between them was 1.02.
Step 2) solving the optimal estimation of beta through the following formula:
wherein F is a basis function matrix;
y is a training data composition vector;
step 3) calculating an output predicted value through the following formula:
in the formula (I), the compound is shown in the specification,
the method is characterized in that a correlation vector of data to be predicted and training data is obtained through a correlation matrix.
The experimental results are shown in fig. 5, where fig. 5 (a) is the prediction results of the present invention and fig. 5 (b) is the prediction results of the general model. In the figure, the abscissa is the real maximum stress obtained by numerical simulation, the ordinate is the maximum stress obtained by modeling according to numerical simulation data, the circle predicts the result, and the dotted line represents that the predicted value is equal to the real value, i.e. when the circle is closer to the dotted line, the more accurate the predicted result is.
As can be seen from fig. 5, compared with the traditional kriging method, the prediction result of the method for modeling the variable capacity data of the combine harvester provided by the invention is more accurate, i.e., the variable capacity data of the combine harvester is modeled more accurately.
The prediction result is further evaluated by adopting an index R formula
In the formula, y i Representing the true value of the ith sample;
n is the number of samples.
Calculated R of the traditional Krigin method 2 0.769, the invention provides an R of the method for modeling the unequal capacity data of the combined harvester 2 Was 0.867. The method for modeling the variable-capacity data of the combined harvester can accurately realize the variable-capacity data modeling of the combined harvester, and the precision is improved by 12.743%.
Example 2
In the embodiment, the modeling of the actually measured unequal-volume data generated in the test process of the combine harvester is realized by using the unequal-volume data modeling method of the combine harvester provided by the invention.
The measured data of the combine harvester operation field used in the embodiment records the measured data including the height of the header, the advancing speed, the rotating speed of the threshing cylinder, the rotating speed of the fan, the opening of the vibrating screen and the cleaning loss number, wherein the sample capacity of the opening of the vibrating screen is 5, and the sample capacity of other parameters is 80.
In this embodiment, a traditional kriging method and the unequal capacity data modeling method for the combine harvester provided by the present invention are respectively adopted to construct an operation performance prediction model, wherein the optimal parameters of the method provided by the present invention are as follows:
[3.69，0.91，2.68，0.19，0.26，0.68，0.61，1.37]。
according to the unequal capacity data modeling method of the combine harvester, provided by the invention, the data x to be predicted are obtained through the following steps new The predicted output value of (a);
step 1) constructing a correlation matrix of a large sample parameter and a small sample parameter based on the optimal parameters;
in the formula, σ s Is the variance corresponding to the small sample parameter, 3.69;
σ l is the variance corresponding to the large sample parameter, which is 2.68;
ρ is the adjustment coefficient between the large and small sample parameters, 1.37.
Step 2) solving the optimal estimation of the beta through the following formula:
wherein F is a basis function matrix;
y is a training data composition vector;
step 3) calculating an output predicted value through the following formula:
in the formula (I), the compound is shown in the specification,
is a correlation vector of the data to be predicted and the training data obtained by the correlation matrix.
The experimental results are shown in fig. 6, where fig. 6 (a) shows the prediction results of the present invention, and fig. 6 (b) shows the prediction results of the general model. In the figure, the abscissa is the true value of the cleaning loss number of 7, and the ordinate is the circle of the predicted value of the cleaning loss numberThe predicted result, the dashed line indicates that the predicted value is equal to the true value, i.e. the more accurate the predicted result is when the circle is closer to the dashed line. As can be seen from the attached figure 6, compared with an extreme learning machine, the prediction result of the unequal capacity data modeling method for the combined harvester provided by the invention is more accurate. Further calculation of R 2 The kriging method is 0.950, the method for modeling the unequal capacity data of the combined harvester is 0.981, the method provided by the invention realizes more accurate modeling of the unequal capacity data of the combined harvester, and the precision is improved by 3.26%.
Example 3
The embodiment provides a combine harvester differential data modeling device which can be integrated on, but not limited to, combine harvester on-board monitoring, remote monitoring, intelligent farms and the like.
The data input module in the device is used for inputting numerical simulation data or actual measurement data of the combine harvester;
a data processing module in the device is used for preprocessing the data of the combine harvester and calibrating a large sample parameter x L Small sample parameter x S And a predicted target parameter y;
data modeling module in the device, based on the large sample parameter x of the depiction L With a small sample parameter x S Constructing an unequal capacity data kriging prediction model according to the correlation relationship, and optimizing parameters of the model to obtain an optimized unequal capacity data kriging prediction model;
in the embodiment, the unequal capacity data modeling device of the combine harvester provided by the invention is integrated on a remote monitoring platform of the combine harvester.
Through the data input module, the combine harvester remote monitoring platform obtains the field measured data of the combine harvester and leads the data into the data processing module.
In the data processing module, input data parameters are calibrated to large sample parameters x L Small sample parameter x S And a predicted target parameter y.
In the data modeling module, according to the large sample parameter x of the drawing L Reference to small sampleNumber x S Constructing an unequal capacity data kriging prediction model according to the correlation relationship, and optimizing parameters of the model to obtain an optimized unequal capacity data kriging prediction model;
and in the data prediction module, predicting the new sample according to the prediction model established by the data modeling module.
The modeling method and the modeling device suitable for the unequal-capacity data of the combine harvester can better realize the design, analysis, optimization and operation and maintenance of the data-driven combine harvester and provide model support for the combination of the data-driven prediction technology and the depth of the combine harvester.
It should be understood that although the specification has been described in terms of various embodiments, not every embodiment includes every single embodiment, and such description is for clarity purposes only, and it will be appreciated by those skilled in the art that the specification as a whole can be combined as appropriate to form additional embodiments as will be apparent to those skilled in the art.
The above-listed detailed description is only a specific description of a possible embodiment of the present invention, and they are not intended to limit the scope of the present invention, and equivalent embodiments or modifications made without departing from the technical spirit of the present invention should be included in the scope of the present invention.
Claims (10)
1. The method for modeling the unequal capacity data of the combined harvester is characterized by comprising the following steps of:
s1, acquiring operation data of the combine harvester;
s2, preprocessing the work data of the combine harvester obtained in the step S1, and calibrating a large sample parameter x L Small sample parameter x S And a predicted target parameter y;
step S3, large sample parameter x of step S2 is described L With a small sample parameter x S Constructing an unequal capacity data kriging prediction model according to the correlation;
and S4, optimizing the parameters of the unequal tolerance data kriging prediction model in the step S3 to obtain the optimized unequal tolerance data kriging prediction model.
2. The method for modeling unequal capacity data of a combine harvester according to claim 1, wherein the means for obtaining combine harvester operation data in step S1 comprises numerical simulation, on-board monitoring, remote monitoring or intelligent farm platform.
3. The method for modeling the unequal capacity data of the combine harvester according to claim 1, wherein the sample parameters and the prediction target parameters in the unequal capacity data kriging prediction model in the step S3 satisfy the following relations:
wherein y is a predicted target parameter;
x is an input variable comprising a large sample parameter x L Small sample parameter x S ，x＝(x L ,x S )；
f j (x) Is the jth basis function;
β j coefficients for the jth basis function;
p is the number of basis functions;
z (x) is a Gaussian process.
4. The method for modeling unequal data of a combine harvester according to claim 3, wherein the gaussian process Z (x) in step S3 satisfies the following condition:
E(Z(x))＝0；
E(Z(x i )Z(x j ))＝σ 2 R(θ,x i ,x j )；
wherein E (-) represents the expectation of the variable;
σ 2 is the sample variance;
R(θ,x i ,x j ) Is a correlation matrix;
theta is a parameter vector of the correlation matrix;
x i is an input vector for the ith data;
x j is the input vector for the jth data.
5. The method of modeling unequal data of a combine harvester according to claim 4, wherein the correlation matrix R (θ, x) i ,x j ) Characterizing a large sample parameter x L And x of small sample parameter S The association relationship between the two is as follows:
in the formula, σ s Is the variance corresponding to the small sample parameter;
σ l is the variance corresponding to the large sample parameter;
ρ is the large sample parameter x L And x of small sample parameter S An adjustment factor therebetween;
p is the number of small sample parameters;
q is the number of large sample parameters.
6. The method for modeling the unequal data of the combine harvester according to claim 5, wherein the step S4 adopts a grayish optimization algorithm to obtain the optimal parameters of the model, and the parameters sigma to be optimized are obtained s 、
σ l 、
Rho is converted into the position of the ith wolf in the species during hunting
Optimizing, wherein the conversion formula is as follows:
7. the method for modeling unequal data of a combine harvester according to claim 1, wherein the step S4 evaluates the training error using Root Mean Square Error (RMSE):
in the formula, n is the number of training samples;
y i is the true output value of the ith training sample;
8. The method for modeling the unequal data of the combine harvester according to claim 6, wherein the grayish optimization comprises the following specific steps:
step 1) setting population scale, maximum iteration times and search space; initializing the population and randomly generating a convergence factor
And a disturbance vector
Assuming that the number of iterations g =1,
a vector of random numbers between 0 and 1,
is a random number between 0 and 1,
linearly decreasing from 2 to 0 as the number of iterations increases;
step 2) calculating and sequencing training errors RMSE corresponding to each individual, and recording the individual position with the minimum training error as the position
Step 3) updating the position of each individual;
step 4) updating the decreasing numerical value
Convergence factor
And disturbance vector
A parameter;
step 5) judging whether the algorithm meets the convergence condition, if so, ending and giving a final result
Namely the optimal model parameters; otherwise, let g = g +1, return to step 2).
9. The method for modeling unequal capacity data of a combine harvester according to claim 5, wherein the optimal unequal capacity data kriging prediction model in the step S4 is as follows:
f(x new ) Denotes x new A vector consisting of values obtained by basis functions;
wherein F is a basis function matrix, F = (F (x) 1 ),f(x 2 ),…,f(x n )) T N is the number of training samples;
F T is the transpose of F;
y is a vector formed by the output of the training data;
10. A modeling device of the unequal capacity data modeling method of the combine harvester according to any one of the claims 1-9, which is characterized by comprising a data input module, a data processing module, a data modeling module and a data prediction module;
the data input module is used for inputting unequal capacity data generated by each operation part of the combine harvester;
the data processing module is used for preprocessing the operation data of the combine harvester and calibrating a large sample parameter x L Small sample parameter x S And a predicted target parameter y;
the data modeling module is used for describing a large sample parameter x L With small sample parameter x S Constructing an unequal capacity data kriging prediction model according to the correlation relationship, and optimizing parameters of the model to obtain an optimized unequal capacity data kriging prediction model;
and the data prediction module is used for predicting a new sample according to the optimized unequal tolerance data Krigin prediction model obtained by the data modeling module.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202211200762.7A CN115495954A (en) | 2022-09-29 | 2022-09-29 | Unequal-volume data modeling method and device for combine harvester |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202211200762.7A CN115495954A (en) | 2022-09-29 | 2022-09-29 | Unequal-volume data modeling method and device for combine harvester |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN115495954A true CN115495954A (en) | 2022-12-20 |
Family
ID=84472489
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202211200762.7A Pending CN115495954A (en) | 2022-09-29 | 2022-09-29 | Unequal-volume data modeling method and device for combine harvester |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN115495954A (en) |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20170177756A1 (en) * | 2015-12-22 | 2017-06-22 | Bwxt Mpower, Inc. | Apparatus and method for safety analysis evaluation with data-driven workflow |
| CN107368649A (en) * | 2017-07-19 | 2017-11-21 | 许昌学院 | A kind of sequence optimisation test design method based on increment Kriging |
| CN115032544A (en) * | 2022-06-30 | 2022-09-09 | 重庆大学 | Method for predicting residual life of power lithium battery under condition of incomplete data |
-
2022
- 2022-09-29 CN CN202211200762.7A patent/CN115495954A/en active Pending
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20170177756A1 (en) * | 2015-12-22 | 2017-06-22 | Bwxt Mpower, Inc. | Apparatus and method for safety analysis evaluation with data-driven workflow |
| CN107368649A (en) * | 2017-07-19 | 2017-11-21 | 许昌学院 | A kind of sequence optimisation test design method based on increment Kriging |
| CN115032544A (en) * | 2022-06-30 | 2022-09-09 | 重庆大学 | Method for predicting residual life of power lithium battery under condition of incomplete data |
Non-Patent Citations (6)
| Title |
|---|
| WEI SUN ET AL.: "Dynamic load prediction of tunnel boring machine (TBM) based on heterogeneous in-situ data", 《AUTOMATION IN CONSTRUCTION》, vol. 92, pages 23 - 34, XP085402198, DOI: 10.1016/j.autcon.2018.03.030 * |
| 刘岩等: "基于天牛须搜索粒子群优化的改进克里金模型", 《国外电子测量技术》, no. 04, pages 6 - 10 * |
| 张宇等: "基于克里金元模型的离心泵水力性能多目标优化", 《华中科技大学学报(自然科学版)》, no. 04, pages 54 - 57 * |
| 李耀辉等: "基于克里金模型的多采样点序列全局优化方法", 《华中科技大学学报(自然科学版)》, no. 12, pages 7 - 11 * |
| 石茂林: "盾构机实测数据建模方法及其应用研究", 《中国优秀博士学位论文全文数据库工程科技Ⅱ辑》, no. 1, pages 15 - 100 * |
| 胡俊峰等: "基于Kriging模型的微夹持器优化设计", 《中国机械工程》, no. 14, pages 1968 - 1975 * |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Haji et al. | Comparison of optimization techniques based on gradient descent algorithm: A review | |
| CN113378939B (en) | Structure digital twin modeling and parameter identification method based on physical driving neural network | |
| CN106650918B (en) | The method and apparatus for constructing system model | |
| CN108537366B (en) | Reservoir scheduling method based on optimal convolution bidimensionalization | |
| CN111915097B (en) | Water quality prediction method for optimizing LSTM neural network based on improved genetic algorithm | |
| KR102688676B1 (en) | An artificial intelligence-based privacy-preserving distribution method for vertically, horizontally and multi-partitioned data and a device thereof | |
| CN115034126A (en) | Method and system for optimizing LSTM neural network model through wolf algorithm | |
| CN114154401A (en) | Soil erosion modulus calculation method and system based on machine learning and observation data | |
| CN114578087B (en) | Wind speed uncertainty measurement method based on non-dominant sorting and stochastic simulation algorithm | |
| CN112966595B (en) | Sensor network data abnormity judgment method based on graph neural network | |
| CN112926251B (en) | Landslide displacement high-precision prediction method based on machine learning | |
| CN111612246B (en) | Method, device, equipment and storage medium for predicting heavy metal content of farmland soil | |
| CN116992341A (en) | Bearing fault diagnosis classification method | |
| CN115495954A (en) | Unequal-volume data modeling method and device for combine harvester | |
| CN117877587A (en) | Deep learning algorithm of whole genome prediction model | |
| CN112416913B (en) | GWO-BP algorithm-based aircraft fuel system state missing value supplementing method | |
| CN115526308A (en) | Time sequence prediction method based on width pruning echo state network | |
| CN114595884A (en) | Genetic intelligent optimization neural network wind power generation equipment temperature prediction method | |
| CN115242428A (en) | Network security situation prediction method based on optimized CW-RNN | |
| Khdhir et al. | Innovative solutions for design and fabrication of deep learning based soft sensor | |
| CN118226751A (en) | Track planning method and system for adjusting posture-changing chassis of crawler-type harvester | |
| CN118072976B (en) | System and method for predicting respiratory tract diseases of children based on data analysis | |
| CN114219068B (en) | Brain effect connection network learning method based on space-time diagram convolution model | |
| CN113035363B (en) | Probability density weighted genetic metabolic disease screening data mixed sampling method | |
| CN115577618B (en) | Construction method and prediction method of high-pressure converter valve hall environmental factor prediction model |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20221220 |